proffesor-for-testing__agentic-qe/docs/analysis/schema-optimization-summary.md

Schema Optimization Analysis - Executive Summary

Date: 2025-12-05 Track: C - Parameter Schema Analysis Issue: #115 - MCP Tool Optimization Swarm: swarm_1764923979903_lbcgvbqu3

---

Key Findings

📊 Optimization Potential

Category	Count	Token Savings
Boolean consolidation candidates	22 tools	~8,000 tokens
Shared type extractions	8 types (40+ usages)	~2,000 tokens
Verbose descriptions	20 descriptions	~500 tokens
Rarely used parameters	5-8 parameters	~800 tokens
TOTAL	60+ tools analyzed	~11,300 tokens (15-20%)

---

Top 5 High-Impact Optimizations

1. Report Generation Tool (5 booleans → 1 enum)

Tool: test_report_comprehensive (Lines 701-721) Current: 5 separate boolean flags Proposed: Single reportDetail enum with 4 levels Savings: ~40% parameter reduction

2. Data Masking Tool (5 booleans → 2 params)

Tool: qe_test_data_mask (Lines 2727-2763) Current: 5 boolean configuration flags Proposed: 1 maskingMode enum + 1 boolean Savings: ~35% parameter reduction

3. Schema Analysis Tool (5 booleans → 1 enum)

Tool: qe_test_data_analyze_schema (Lines 2789-2809) Current: 5 analysis scope booleans Proposed: Single analysisScope enum Savings: ~45% parameter reduction

4. Shared Priority Type (10 tools)

Type: PriorityLevel enum Current: Duplicated 10+ times Proposed: Extract to shared type Savings: ~50% on priority parameters

5. Shared Report Format (8 tools)

Type: ReportFormat enum Current: Duplicated 8 times Proposed: Extract to shared type Savings: ~40% on format parameters

---

Implementation Roadmap

Phase 1: Boolean Consolidation (Week 1)

• test_report_comprehensive (5 → 1)
• qe_test_data_analyze_schema (5 → 1)
• qe_test_data_mask (5 → 2)
• test_coverage_detailed (4 → 1)
• coverage_analyze_with_risk_scoring (4 → 1)

Deliverables: New enum parameters, backward compatibility layer

Phase 2: Shared Types (Week 2)

• Create src/mcp/schema-types.ts
• Extract 8 common types
• Update 40+ tool usages
• Add JSDoc documentation

Deliverables: Shared types module, migration guide

Phase 3: Description Cleanup (Week 3)

• Shorten 20 verbose descriptions
• Remove redundant prefixes
• Standardize format

Deliverables: Cleaner descriptions, updated docs

Phase 4: Deprecation (Week 4)

• Mark rare parameters deprecated
• Add migration warnings
• Plan v3.0 breaking changes

Deliverables: Deprecation notices, v3.0 plan

---

Tools with Most Optimization Potential

Tool	Boolean Count	Line Range	Impact
qe_test_data_mask	5	2727-2763	Very High
qe_test_data_analyze_schema	5	2789-2809	Very High
test_report_comprehensive	5	701-721	Very High
test_coverage_detailed	4	774-793	High
coverage_analyze_with_risk_scoring	4	1897-1900	High
flaky_analyze_patterns	4	2081-2084	High
performance_analyze_bottlenecks	4	2146-2149	High
qe_requirements_generate_bdd	4	3048-3063	High

---

Common Shared Types to Extract

1. PriorityLevel (10 usages) - Lines: 429, 755, 996, 1033, 1338, 1955, 2296, 2943, 3005, 3177
2. ReportFormat (8 usages) - Lines: 264, 692, 2165, 2334, 3435, 3747
3. FleetIdParam (8 usages) - Lines: 160, 271, 395, 467, 1817, 3248
4. AgentIdParam (6 usages) - Lines: 216, 860, 999, 3282, 3810, 3847
5. TestFramework (6 usages) - Lines: 105-108, 203-206, 547-550, 2110-2113
6. DetailLevel (5 usages) - Lines: 768-772, 1548-1552, 3590-3594
7. Environment (7 usages) - Lines: 100-103, 239-242, 3629, 3672
8. QualityScope (4 usages) - Lines: 289-292, 339-342, 763-766

---

Verbose Descriptions to Shorten

Line	Current Length	Proposed Length	Savings
3806	146 chars	56 chars	62%
3843	136 chars	36 chars	74%
3879	146 chars	35 chars	76%
3926	146 chars	39 chars	73%
1773	101 chars	39 chars	61%
1833	88 chars	32 chars	64%

---

Rarely Used Parameters (Removal Candidates)

1. collectCoverage in test_execute_parallel (Line 615-618)

   • Reason: Redundant with dedicated coverage tools

2. predictFailures in test_optimize_sublinear (Line 658-661)

   • Reason: Separate concern, dedicated tool exists

3. includeMetadata in memory_retrieve (Line 855-859)

   • Reason: Rarely needed, advanced use case

4. dryRun in workflow_execute (Line 1229-1232)

   • Reason: Should be separate validation tool

5. historicalDataDays in predict_defects (Line 356-362)

   • Reason: Over-specified, use enum instead

---

Example: Before/After

Report Generation (62% token reduction)

// BEFORE (120 tokens)
includeCharts: { type: 'boolean', default: false, description: 'Include visual charts' }
includeTrends: { type: 'boolean', default: false, description: 'Include trend analysis' }
includeSummary: { type: 'boolean', default: true, description: 'Include summary section' }
includeDetails: { type: 'boolean', default: false, description: 'Include detailed test information' }
structured: { type: 'boolean', default: true, description: 'Use structured output (for JSON)' }

// AFTER (45 tokens)
reportDetail: {
  type: 'string',
  enum: ['minimal', 'summary', 'detailed', 'comprehensive'],
  default: 'summary'
}

Shared Types (50% token reduction per usage)

// BEFORE (30 tokens per usage, 10 usages = 300 tokens)
priority: {
  type: 'string',
  enum: ['low', 'medium', 'high', 'critical'],
  default: 'medium',
  description: 'Task priority level'
}

// AFTER (15 tokens per usage, 10 usages = 150 tokens)
priority: PriorityParam

---

Risk Assessment

Low Risk

• Shared type extraction (backward compatible)
• Description shortening (no API change)

Medium Risk

• Boolean consolidation (requires migration layer)
• New enum parameters (needs testing)

High Risk

• Removing deprecated parameters (breaking change)
• Requires v3.0 major version

---

Success Metrics

• 15-20% reduction in schema token overhead
• 65% fewer boolean parameters
• 8 reusable shared types
• 100% backward compatibility in v2.x
• Complete migration guide
• All tests passing

---

Next Steps

1. Review this analysis with team
2. Prioritize optimizations by impact
3. Create implementation PRs
4. Test backward compatibility
5. Document migration path
6. Deploy in phases (v2.2, v2.3, v3.0)

---

Full Report: schema-optimization-report.md Generated by: Track C Analysis Agent Coordination: Part of 4-track parallel optimization swarm